RESEARCH RESOURCE

PROTEOMICS A systems-wide screen identifies substrates of the SCFbTrCP ligase

Teck Yew Low,1,2 Mao Peng,1,2* Roberto Magliozzi,3* Shabaz Mohammed,1,2† Daniele Guardavaccaro,3 Albert J. R. Heck1,2‡

Cellular are degraded by the ubiquitin-proteasome system (UPS) in a precise and timely fashion. Such precision is conferred by the high substrate specificity of ubiquitin ligases. Identification of substrates of ubiquitin ligases is crucial not only to unravel the molecular mechanisms by which the UPS controls degradation but also for drug discovery purposes because many established UPS substrates are implicated in disease. We developed a combined bioinformatics and affinity purification– mass spectrometry (AP-MS) workflow for the system-wide identification of substrates of SCFbTrCP,a member of the SCF family of ubiquitin ligases. These ubiquitin ligases are characterized by a multi- subunit architecture typically consisting of the invariable subunits Rbx1, Cul1, and Skp1, and one of 69 F-box proteins. The F-box protein of this member of the family is bTrCP. SCFbTrCP binds, through the WD40 b repeats of TrCP, to the DpSGXX(X)pS diphosphorylated motif in its substrates. We recovered 27 pre- Downloaded from viously reported SCFbTrCP substrates, of which 22 were verified by two independent statistical proto- cols, thereby confirming the reliability of this approach. In addition to known substrates, we identified 221 proteins that contained the DpSGXX(X)pS motif and also interacted specifically with the WD40 repeats of bTrCP. Thus, with SCFbTrCP, as the example, we showed that integration of structural infor- mation, AP-MS, and degron motif mining constitutes an effective method to screen for substrates of ubiquitin ligases. http://stke.sciencemag.org/

INTRODUCTION example, an E3 can be perturbed by RNAi (RNA inter- Relative to most other protein posttranslational modifications (PTMs), ference), chemical inhibitors, or genetic manipulations, whereafter the per- ubiquitylation is unique because ubiquitin can be conjugated to a substrate turbed samples and controls may be analyzed by quantitative mass protein in various chain linkages through which the modified substrates spectrometry (MS) for any change in protein abundance or the amount are sorted to different cellular fates (1). Substrates polyubiquitylated with of PTM-modified proteins (8–10). The availability of an antibody that rec- Lys48 linkages are recognized and degraded by the 26S proteasome (2). ognizes diglycine residues, derived from ubiquitin after trypsin cleavage

Target substrates are modified by the sequential actions of a ubiquitin- (11), enables the isolation and identification by MS of ubiquitylated pep- on May 9, 2018 activating enzyme (E1), a ubiquitin-conjugating enzyme (E2), and a ubiqui- tides, thereby providing information about the site of ubiquitylation to tin ligase (E3). The huge repertoire of more than 700 E3 ubiquitin ligases complement quantitative assays (12, 13). in mammals provides the specificity required for the recognition of diverse Alternative approaches based on affinity purification followed by mass substrates (3, 4). Because numerous substrates targeted by the ubiquitin- spectrometry (AP-MS) (14–16) using epitope tagging are also feasible for proteasome system (UPS) are oncoproteins or tumor suppressors, which such screens. In AP-MS methods, a protein of interest (bait) is fused in are potential targets for therapy (5, 6), it is crucial to develop high- frame with an epitope tag, which enables the isolation of the protein com- throughput assays to identify the complete repertoire of substrates of spe- plexes (preys). The affinity-purified complexes are then identified by MS. cific E3 ubiquitin ligases. Although AP-MS experiments have contributed to the identification of Because E3 ubiquitin ligases, as well as other modifying enzymes, are substrates of monomeric E3s, identification of substrates of multisubunit believed to interact transiently with their substrates in a “kiss-and-run” E3s is hampered by the weak affinity of substrates for the subunit of the manner (7), researchers tend to assay the products of the enzymatic activ- E3 that recognizes the substrate and because the complex between the sub- ity rather than the enzyme-substrate interactions. To this end, quantitative strate and the recognition subunit is transient (17). proteomics methods are widely used because they allow the identification Here, we used a stringent AP-MS approach to systematically identify and quantitation of protein substrates with relatively high throughput. For substrates of SCFbTrCP, a ubiquitin ligase that targets key regulators of cell proliferation, cell growth (15), and apoptosis (18) for proteasomal degra- dation. SCFbTrCP consists of Skp1, Cul1, Rbx1, and the F-box protein 1Biomolecular Mass Spectrometry and Proteomics, Bijvoet Center for Biomo- bTrCP that acts as a substrate recognition subunit (Fig. 1). Mammals have lecular Research and Utrecht Institute for Pharmaceutical Sciences, Utrecht Uni- two paralogs of bTrCP: bTrCP1 (also known as FBXW1) and bTrCP2 versity, Padualaan 8, 3584 CH Utrecht, Netherlands. 2Netherlands Proteomics 3 – (also known as FBXW11); however, their biochemical properties appear Center, Padualaan 8, 3584 CH Utrecht, Netherlands. Hubrecht Institute KNAW bTrCP (Royal Netherlands Academy of Arts and Sciences) and University Medical indistinguishable. Although more than 50 substrates of SCF have Center Utrecht, Uppsalalaan 8, 3584 CT Utrecht, Netherlands. been identified since 1999, the list is undoubtedly incomplete. Certain *These authors contributed equally to this work. E3 ligases of the SCF family recognize their substrates through a phos- †Present address: Departments of Chemistry and Biochemistry, University of bTrCP Oxford, New Biochemistry Building, South Parks Road, Oxford OX1 3QU, phorylated motif called a phosphodegron. Most known SCF sub- UK. strates share the canonical phosphodegron DpSGXX(X)pS, which is a ‡Corresponding author. E-mail: [email protected] diphosphorylated motif that mediates the interaction of the substrate with

www.SCIENCESIGNALING.org 16 December 2014 Vol 7 Issue 356 rs8 1 RESEARCH RESOURCE

Ub Ub These sequences included 45 of the 47 reported phosphodegrons used for Phosphodegron Ub Ub Ub generating the consensus sequence; the two that were not recovered were Ub Ub viral protein U (Vpu) and the RNA binding protein ELAV1. Vpu is a viral Substrate DSGXX(X)S P P protein encoded by HIV-1. Because the human Swiss-Prot database that we used for FIMO analysis does not contain this viral protein, the phos- F-box βTrCP Substrate receptor phodegron for Vpu (RAEDSGNESE) was not recovered. Our recalcula- E2 − Adaptor SKP1 WD40 repeats P 6 RBX1 Ub tion shows that the FIMO value for this phosphodegron is 1.9 × 10 . The phosphodegron sequence (TNYEEAAMAI) of ELAV1 is so devi- Scaffold CUL1 N ated from the consensus that the FIMO P value is 0.00124, lower than the − E1 Ub P ≤ 1×103 cutoff, and hence is also not included. Ub Because FIMO P valuescanbeusedtoestimatehowsimilareachpu- Fig. 1. The architecture of wild-type SCFbTrCP. SCFbTrCP consists of the core tative phosphodegron is to the consensus motif, we ranked each matched subunits Cul1, Rbx1, and Skp1, and the F-box protein bTrCP that deter- motif according to its respective FIMO P values (lowest first, represent- mines substrate specificity. Cul1 is a scaffold protein that interacts through ing the closest match to the consensus motif). As expected, the top-ranking its N terminus with the adaptor protein Skp1 and through its C terminus with reported phosphodegrons generally contained motifs that matched more the RING finger protein Rbx1, which in turn binds to a specific ubiquitin- closely to the DpSGXX(X)pS motif, for example, eEF2K (ranked 1), CTNNB1 conjugating enzyme (E2). Skp1 interacts with the F-box domain of bTrCP, (ranked 11), DLG1 (ranked 246), and SIPA1L1 (ranked 404). For ease of which, in turn, recruits substrates through its WD40 repeats. These repeats reference, we named them category I motifs. Following the SIPA1L1, we form a b-propeller structure, which specifically recognizes a diphosphory- observed a sharp drop in the ranking for motifs derived from substrates lated motif with the consensus DpSGXX(X)pS, known as phosphodegron. such as PLK4 (ranked 1214), MYC (ranked 1426), CDC25B (ranked Downloaded from 2931), TWST1 (ranked 3778), and MCL1 (ranked 5218), which have less canonical motifs. The second group is termed category II motifs. the WD40 repeats of bTrCP (Fig. 1) (19, 20). Mutation of either the phos- To further refine and corroborate our motif analysis, we queried the phodegron or the WD40 repeats abolishes substrate binding to bTrCP (21). PhosphoSitePlus repository (35), which contains phosphoproteomics and The phosphodegrons of all reported SCFbTrCP substrates are conserved ubiquitomics data. In total, we found 3881 phosphorylation sites reported in vertebrates (22). for our putative degrons. This analysis revealed that 46% (3507 of 7623) http://stke.sciencemag.org/ A challenge in AP-MS analyses is discriminating bona fide interacting of the proteins with putative degrons have at least one ubiquitylation site partners from nonspecific protein contaminants (23). To resolve this pro- (table S2). However, because of the low abundance of some modifications blem, we used label-free quantitative proteomics in combination with two at specific sites and limitation of proteomics technology, we failed to find different background-filtering procedures. First, we used the CRAPome matching phosphoproteomics data for some well-characterized SCFbTrCP (24) software suite, which not only includes a large list of common nonspe- substrates, such as PER1, WEE1, EPOR, CPEB-1, PLK4, and FANCM. cific contaminants but also provides, due to the embedded SAINT algo- Evolutionary conservation can be used to distinguish between functional rithm, a standardized scoring scheme for ranking each prey candidate and nonfunctional motifs (36–38), and thus, we subjected each bTrCP against these contaminants (23, 25, 26). We also analyzed the data with degron to a conservation analysis. Among the 45 known degrons identi-

a second strategy based on precursor ion (MS1) intensity using the MaxQuant- fied in our experiment, 42 degrons had a conservation score ≥0.7 between on May 9, 2018 Perseus software suite (27–30), which is another method that enriches for human and mouse orthologs (table S2). However, only 23 degrons had a high-confidence interactors using label-free quantitative proteomics. By conservation score ≥0.7 between human and fish, and only 1 degron analyzing the AP-MS data with these software tools, we discriminated (SYLDSGIHSG) for CTNNB1 was conserved between human and Dro- specific protein interactors from presumed background contaminants sophila. This analysis indicated that bTrCP mainly functions in vertebrates, (31) and studied dynamically regulated interactions (32–34). Using this although its ortholog (called Slimb) also exists in flies. By setting a con- workflow, we identified most of the known substrates of SCFbTrCP as well servation score of 0.7 as observed in human and mouse orthologs, we re- as many putative previously unknown substrates of this ubiquitin ligase. duced the total 17,779 putative degrons to 13,109 (74%). We validated this approach by examining several putative SCFbTrCP sub- strates by immunoprecipitation followed by immunoblotting and in vitro ubi- Identification of known and candidate quitylation assay. We envisage that this approach may be useful for substrates of bTrCP2 identifying substrates of other E3 ubiquitin ligases. For the AP-MS experiments, we designed three different constructs encod- ing bTrCP2 carrying N-terminal 2xFLAG and 2xHA epitope tags (Fig. 3, A to C). The rationale of using two sequential immunoprecipitations with RESULTS two different epitope tags is to reduce the amount of background proteins that copurify, albeit at the expense of weaker protein-protein interactions Mining bTrCP substrate phosphodegron motifs in the (PPIs), which may be lost due to the high stringency inherent to this se- human proteome quential method. We used the epitope tags without any bTrCP2-based fu- We first extracted 47 phosphodegrons from reported substrates of bTrCP sion protein (EV) as one negative control. The three baits consist of the (table S1) and used the MEME program to formulate a stringent consen- wild-type bTrCP2 (Fig. 3A), bTrCP2 (R447A) (Fig. 3B), and bTrCP2-DF sus motif (Fig. 2A). We then probed the human proteome (Swiss-Prot, (Fig. 3C). The Arg-to-Ala mutation in bTrCP2 (R447A) is within the human, July 2013, 20,277 entries) using this consensus motif with the WD40 b-propeller domain and abrogates the interaction with and conse- FIMO program, which generates a P valueforeverymotifmatched(the quently ubiquitylation of SCFbTrCP substrates (19, 21); however, like the smaller the P value, the more closely the motif matches the consensus). corresponding bTrCP1 (R474A) mutant, bTrCP2 (R447A) retains the Overall, we obtained 17,779 short sequences that met the requirement ability to interact with other components and regulators of the SCF com- of a P ≤ 1×10−3 and represented 7623 proteins (table S2 and Fig. 2B). plex (21, 39, 40). We assumed that such a single amino acid mutant serves

www.SCIENCESIGNALING.org 16 December 2014 Vol 7 Issue 356 rs8 2 RESEARCH RESOURCE

A We identified 1626 proteins in the AP- 4 MS experiments, 27 of which were previous- ly reported substrates of SCFbTrCP (tables S3 and S4). Because the bottleneck for AP-MS 3 experiments is the selection of bona fide pro- tein binders from an abundance of back- ground contaminants, we predicted that most of the 1626 proteins were nonspecific or sec- Bits 2 ondary binders. To differentiate genuine PPIs from background contaminants, we used two distinctive quantitative proteomics filtering 1 — S Y tools the CRAPome and Perseus software TS I GT — NDE LSS suites and compared immunopurified sam- S PTS P E FHLE PVS Q R RE G I KY L ANDDG W T D S F T C ples with negative controls that did not con- E C V D Q R NM LN A F V A D S I P ED E 0 GGL A tain any bait (Fig. 3D). 1 2 3 4 S5 G6 7 8 9 10 The CRAPome software enables back- B ground filtering using spectra count data acquired by AP-MS (24). It consists of a database of categorized background con- Reported phosphodegron taminants with two scoring schemes to Downloaded from eEF2K New phosphodegron independently rank each pair of PPI identi- PRLR SNAI1 fied by immunoprecipitation. Here, we ATF4 UHRF1 ranked wild-type bTrCP2, bTrCP2 (R447A), CTNNB1 and bTrCP2-DF individually against the

Canonical RASSF1-c FBXO5 negative control (EV). For each bait-prey in- IKBB

CDC25A http://stke.sciencemag.org/ FANCM PDCD4 teraction, CRAPome computed a probability- YAP1 CLSPN NFKBIE NFE2L2 based SAINT score (23) and a separate fold PER1 BCL2L11 REST FGD1 PER2 change (FC) score (24) (table S3A). For USP37 GHR DEPTOR IFNAR1 Vpu Category I SP1 IKBA WEE1 the wild-type bTrCP2 immunoprecipitation, BORA WWTR1 EPOR NFKB1 20 of the 27 reported substrates had high DLG1 NFKB2 (FIMO P value) CPEB1−a SAINT scores of ≥0.90 (fig. S1 and table 10 SIPA1L1 S3B), suggesting that the SAINT score alone Category II may be sufficient to identify genuine binders. −Log Therefore, we filtered all PPIs from wild-type

bTrCP2, bTrCP2 (R447A), and bTrCP2-DF on May 9, 2018 immunoprecipitations against EV with a SAINT score ≥0.90 (fig. S1). However, this

Noncanonical analysis revealed that even specific binders

345678 of bTrCP2arenotalwaysbonafidesub- strates, as exemplified by the immunopre- cipitation of SCF subunits Skp1, Cul1, and 0 5000 10,000 15,000 20,000 Rbx1, which all had SAINT scores ≥0.9. Sequences matched to consensus motif The FC score not only estimates the re- Fig. 2. Mining the for the DpSGXX(X)pS motifs. (A) Consensus motif derived from the lative abundance (FC) of proteins between phosphodegrons from reported bTrCP substrates: 47 well-characterized phosphodegrons were used two samples but also has a wider dynamic to build a consensus motif using MEME (http://meme.nbcr.net), highlighting the already known canonical range than the SAINT score (fig. S1). There- DpSGXX(X)pS sequence motif where D, S, and G are the best-conserved amino acid residues. (B)Dis- fore, we computed a second value (FC ratio) tribution of FIMO −log P values against all possible sequence motifs in the human proteome: all 17,779 from the FC scores of wild-type bTrCP2 sequence motifs obtained from FIMO analysis of the human proteome are each plotted against their re- and bTrCP2-DF against those obtained for spective −log P values. Some proteins have more than one motif, for example, WEE1 has both a category I bTrCP2 (R447A), which should not bind and a category II motif. any substrates. The FC ratio provides a de- gree of confidence that the bound proteins are substrates, as well as providing infor- as a critical negative control in the context of substrate screening, not only mation that can be used to estimate binding stoichiometry. By setting a because it allows selection against background contaminants but also be- SAINT score cutoff of 0.90 and an FC ratio cutoff of 10.00, the SCF sub- cause it enables the identification of interactors that are specific for the units bTrCP2, Skp1, Cul1, and Rbx1 (SAINT score ≥0.90) were classified WD40 domain. The bTrCP2-DF mutant carries a deletion in the F-box as nonsubstrates, whereas 18 of the 27 reported substrates were retained motif and, when overexpressed in cells, acts as a dominant-negative mu- (Fig. 4A). Collectively, from the data, we defined 338 specific interactors tant (41, 42) because it binds substrates but cannot interact with Skp1, of the WD40 repeat region of bTrCP2 [154 for wild-type bTrCP2 and 231 Cul1, Rbx1, and the E2 enzyme. for bTrCP2-DFagainstbTrCP2 (R447A), with 47 in common] on the

www.SCIENCESIGNALING.org 16 December 2014 Vol 7 Issue 356 rs8 3 RESEARCH RESOURCE

Ub Ub A Ub B C Ub DSGXX(X)S Ub Ub P Ub P DSGXX(X)S DSGXX(X)S P P R447A P P β FLAG HA F-box TrCP FLAG HA F-box βTrCP FLAG HA βTrCP SKP1 WD40 repeats E2 SKP1 WD40 repeats E2 RBX1 Ub RBX1 Ub WD40 repeats CUL1 N CUL1 N

E1 Ub E1 Ub Ub Ub βTrCP WT βTrCP (R447A) βTrCP F

D βTrCP IP

Known βTrCP LC-MS/MS substrates Downloaded from Degron consensus Human MaxQuant motifs extraction proteome

Motif search Fig. 3. bTrCP constructs and work- FIMO Proteome flow for analysis. (A) Wild-type (WT) bTrCP with two HA (hemagglutinin) http://stke.sciencemag.org/ and two FLAG tags. (B) bTrCP R447A Evolutionary mutant that cannot bind substrates. conservation CRAPome Perseus (C) bTrCP-DF mutant lacking the F-box and that can bind substrate but cannot Datasets Substrate interact with the other components of overlapping prediction the SCF complex. (D)Overviewofthe experimental and bioinformatics work- Checking flow. bTrCP IP represents the pro- PTM database Putative new teins coimmunoprecipitated after sequential immunoprecipitation with

substrates on May 9, 2018 HA and FLAG tags from cells expressing the WT bTrCP or one of the mu- tant constructs. Experimental validation

basis of the CRAPome analysis (table S3C), including 20 (18 from wild Reproducibility of AP-MS experiments type and 3 from DF) of 27 previously reported substrates. We have been working toward the goal of identifying substrates of SCFbTrCP To increase confidence in the identified substrates, we performed a for several years and earlier gathered data using an identical sequential im- second filtering analysis by Perseus (43) using the normalized MS1 munoprecipitation AP protocol in conjunction with older versions of the [label-free quantification (LFQ)] intensity (44) obtained from MaxQuant. separation, MS, and bioinformatics technologies. These older data led to For each pair of interactions, we performed Student’s t tests (table the identification, validation, and characterization of four SCFbTrCP sub- S4A and Fig. 4B). With the Perseus workflow, we statistically confirmed strates: eEF2K (21), RAPGEF2 (39), TFAP4 (40), and bHLH40e (45). 20 of the 27 reported substrates (table S4B), of which 18 were signif- To evaluate the reproducibility of our AP-MS workflow, we reanalyzed icantly enriched in the wild-type bTrCP2 immunoprecipitation data and these older data (data set 2) using CRAPome (table S5). Because these 7inthebTrCP2-DF immunoprecipitation data. Together, we obtained older experiments were performed with MS instruments that are at least 317 significant putative hits (163 for wild-type bTrCP2 and 221 for two generations older (Orbitraps Classic and Discovery) and are less sen- bTrCP2-DF, with 61 in common between the two) using label-free quan- sitive and lower in sequencing speed compared to the Orbitrap Elite, fewer titation (table S4C). proteins were identified (725 in data set 2 versus 1626 in data set 1) (table Overall, we identified 338 bTrCP2 substrate candidates that were S6). Of the 725 proteins identified, 531 overlap with data set 1 (73%), in- validated with the CRAPome and 317 validated with Perseus; the candi- dicating that the AP-MS experiments are consistent and reproducible. dates validated by both methods represented 207 proteins (Fig. 4C). Among Upon the CRAPome analysis, we assigned 47 proteins from data set 2 the 27 previously reported substrates, we confirmed 22 in the CRAPome as putative bTrCP substrates, of which 36 overlapped with the 154 puta- analysis (20 substrates) and Perseus (20 substrates), and the overlap between tive substrates in data set 1. The lower number of putative substrates val- them is 18 (Fig. 4D). idated by the CRAPome analysis in data set 2 reflected the lower spectra

www.SCIENCESIGNALING.org 16 December 2014 Vol 7 Issue 356 rs8 4 RESEARCH RESOURCE

A T C C L C CTNNB1 CD CDC25A PE PER1 US USP37 CLSPN C HNRNPD UH UHRF1 CD CDC25B eEF2K eEF PD p53p PDCD4 WEE1 1.0 WEE W REST RES

PER2 PE P Perseus E Members of S 0.9 SCF E3 ligase 110207 131 complex 0.8 CRAPome 0.7 EMI1 0.6 RASSF1

0.5 D 0.4

SAINT score WT SAINT score Perseus BORA 0.3 2218

0.2 CRAPome

0.1 Downloaded from 0.0

4 0 4 8 12 16 2024 3228 36 40 44 48 52 E

FC score ratio WT/R447A Perseus B 42123 56 http://stke.sciencemag.org/ 8.0 CRAPome

7.0 USP37 Fig. 4. Validation of substrate candidates. 6.0 (A) Validation by statistical analysis with CTNNB1 HNRNPD CRAPome. The distribution of the SAINT ATF4 CLSPN 5.0 ELAV1 CDC25A score and the FC score ratio (FC ratio) for SKP1 PDCD4 NFE2L2 the WT bTrCP and empty vector (EV) pair NF-KB2 eEF2K versus the bTrCP (R447A) and EV pair are on May 9, 2018 4.0 IRAK1 UHRF1 CDC25B p53 shown. The SCF subunits bTrCP2, SKP1, WEE1 REST RASSF1 RBX1, and CUL1 are indicated with aster- 3.0 CUL1 value) WT versus ( P value) isks. (B) Validation by statistical analysis

10 with Perseus. A volcano plot from label- PER1 2.0 RBX1 free quantitative proteomics analysis of FBXW11 –Log bTrCP WT versus bTrCP (R447A) samples 1.0 t NF-KB1 is shown. Logarithmized ratios ( test dif- ference) are plotted against the negative 0.0 logarithmic P value of the t test. Back- –8 –6 –4 –2 0 2 4 6 8 10 12 14 ground binding proteins have a ratio close T test difference WT versus R447A to 1:1 and are located close to the vertical 0-line (gray dots). Red dots on the right of vertical 0-line indicate significant hits for substrates because they are more enriched in the WT sample compared to the R447A mutant. Red triangles represent reported substrates; gray triangles represent previously reported substrates that are not statistically significant in our AP-MS experiments; as- terisks represent SCF subunits. (C) Venn diagram of the 338 candidates verified by the CRAPome analysis and 317 verified by Perseus analysis. (D) Venn diagram of the reported substrates verified by CRAPome and Perseus analysis methods. (E) Venn diagram of the substrates analyzed with CRAPome or Perseus but limited to those from the AP-MS data that also contained a degron motif (FIMO P ≤ 1×10−3) and a conservation score of at least 0.7. counts (due to the lower sensitivity and speed of older MS), which af- FC ratio ≥10), only 13 from data set 2 met the criteria for putative sub- fected calculation of the multiple CRAPome scores. strates, of which 12 overlapped with those in data set 1. Nevertheless, eEF2K We also examined the number of reported substrates recovered. We (21), RAPGEF2 (39), TFAP4 (40), and bHLH40e (45) met the stringent identified 25 previously reported bTrCP substrates from data set 2 (table criteria for putative substrates in both data set 1 and data set 2 after re- S6), of which 21 were also present in the 27 identified in data set 1. How- analysis. Therefore, the AP-MS experiments were reproducible, but illus- ever, when we applied the stringent CRAPome filters (SAINT score ≥0.9, trate how the continuous advancement in sensitivity and sequencing speed

www.SCIENCESIGNALING.org 16 December 2014 Vol 7 Issue 356 rs8 5 RESEARCH RESOURCE of MS also leads to a parallel and inevitable improvement in MS data in borderline of the category I motifs (Fig. 2B). Therefore, this analysis sug- terms of both quality and quantity, which allows more stringent data gested that the putative substrates within bins 1 to 4 are high-confidence filtering and the recovery of more confident hits. SCFbTrCP substrates, which is consistent with the recent characterization of eEF2K as an SCFbTrCP (21). Integration of AP-MS and bioinformatics The fifth and sixth bins contain the highest number of candidates (54 prediction data and 134, respectively) but also the lowest number of established substrates Taking the 448 specific interactors classified by each CRAPome and (1 and 3, respectively). Most candidates in these bins have motifs (cat- Perseus analyses as input, we filtered these proteins by the presence of egory II motifs) that diverge from the canonical phosphodegron motif motifs matching the phosphodegron consensus sequence and by the evo- of the known substrates. Although we considered the candidates with cat- lutionary conservation of the motifs. We combined the AP-MS data with egory II motifs less confident, some candidates from these bins have been the bioinformatics predictions by first requiring all putative substrates to identified as SCFbTrCP substrates, such as TFAP4 (40)(fifthbin),bHLHe40 haveadegronmotif(FIMOP ≤ 1×10−3) and a conservation score of at (45) (fifth bin), and RAPGEF2 (39) (sixth bin). least 0.7 (table S7A). These additional criteria reduced the number of pu- tative substrates in the combined CRAPome and Perseus data set from 448 Additional biochemical analysis of previously unknown to 221 (Fig. 4E and table S7A). bTrCP substrates Next, we divided these 221 putative substrates into six bins according To provide additional support that the previously unknown interactors to their FIMO P values (Fig. 5 and table S7B). The number of putative of bTrCP that we identified are substrates, we performed additional substrates increased with the FIMO P value bins. Second, although con- biochemical analysis of two of the putative substrates of SCFbTrCP. sisting of smaller numbers, bins with lower P values are more enriched in We selected for additional biochemical analysis RAPGEF2, a guanine reported substrates. For example, in the first (lowest) bin, eEF2K is the nucleotide exchange factor for the guanosine triphosphatase (GTPase) Downloaded from only candidate in the bin and is a reported substrate (enrichment of re- RAP, which is involved in the development and maintenance of epithe- ported substrates is 100%). This is followed by the second and third bins, lia (46–48), and TBC1D4, a Rab GTPase-activating protein, which is both displaying 67% enrichment (two of three and six of nine candidates involved in insulin-regulated trafficking of the glucose transporter GLUT4 in the bin represent reported substrates, respectively). In the final bin con- (49, 50). taining proteins with category I motifs, there is 35% enrichment. Arranging To test the interaction of bTrCP with RAPGEF2 and TBC1D4, we im- these candidates according to the ranks of their predicted degron motifs munoprecipitated from human embryonic kidney (HEK) 293T cells http://stke.sciencemag.org/ (table S5B) placed DOCK7 as the last candidate in the fourth bin, and this FLAG-tagged bTrCP1, FLAG-tagged bTrCP2, several other FLAG-tagged protein had a phosphodegron motif that was ranked 432, which is at the F-box proteins, as well as CDH1 and CDC20, which are two WD40- containing substrate recognition subunits of the APC/C (anaphase-promoting complex/cyclosome) ubiquitin ligase. After immunoblotting, we observed Category II motifs that bTrCP1 and bTrCP2, and not the other F-box proteins, coimmuno- 140 3 precipitated endogenous RAPGEF2 and TBC1D4 (Fig. 6A), as well as other previously reported substrates, REST and b-catenin (also known 120 as CTNNB1), which served as positive controls. 447 474

The Arg residue located within the WD40 repeats of bTrCP2 (Arg on May 9, 2018 in bTrCP1) is essential for the bTrCP-substrate interaction (21, 39). To assess 100 whether the observed interactions depended on the WD40 b-propeller structure, we immunoprecipitated the FLAG-tagged wild-type bTrCP2 80 and the bTrCP2 (R447A) mutant and examined their binding to these sub- strates. As expected, wild-type bTrCP2, but not the bTrCP2 (R447A) mu- 134 tant, immunoprecipitated endogenous RAPGEF2, bHLH40e (also known 60 as DEC1), and TBC1D4 (Fig. 6B), as well as the known SCFbTrCP sub- 1 strates PDCD4 and b-catenin. Category I motifs bTrCP 40 If RAPGEF2 and TBC1D4 are bona fide substrates of SCF ,they should be ubiquitylated by this ubiquitin ligase. Purified wild-type bTrCP, Counts of putative substrates of putative Counts 54 but not the inactive bTrCP-DF mutant, induced the ubiquitylation of 20 7 TBC1D4 (Fig. 6C) and RAPGEF2 (39) when assayed in vitro with other 6 subunits of the SCF complex. Thus, these data indicated that RAPGEF2 20 bTrCP 2 9 and TBC1D4 can be ubiquitylated by SCF . 0 1 3 –8 –7 –6 –5 –4 –3 1.00 × 10 1.00 × 10 1.00 × 10 1.00 × 10 1.00 × 10 1.00 × 10 Identification of putative regulators or nonsubstrate FIMO P values interactors of SCFbTrCP In addition to identifying putative substrates of bTrCP, our CRAPome and Putative substrates Reported substrates Perseus analyses also indicated that proteins that are not substrates of Fig. 5. Distribution of putative substrates by binned FIMO P values. Dividing bTrCP but instead interact with this F-box protein were also enriched in the 221 candidates into six bins according to their FIMO P values. All can- the AP-MS analysis. For example, the SCF core subunits Skp1 and Cul1 didates have a least one putative degron (FIMO score P ≤ 1×10−3)witha were enriched, but these are not substrates of bTrCP. The interaction of conservation score ≥0.7. Green represents number of candidates within a Skp1 and Cul1 with bTrCP was readily detected by coimmunoprecipi- bin that are already reported. The lower number in each bin represents the tation from HEK293T cells expressing FLAG-tagged wild-type bTrCP2 total number of candidates in the bin. or bTrCP2 (R447A). We found that HUWE1 and UBR5 (also known as

www.SCIENCESIGNALING.org 16 December 2014 Vol 7 Issue 356 rs8 6 RESEARCH RESOURCE

B IP anti-FLAG A IP anti-FLAG T W

7 5 8 1 2 W rCP rCP DC20 V BXW4 BXW TrCP2 T T TrCP2 (R447A) FBX FBXW β FBXW2 C F E β F CDH1 EV β β

TBC1D4 TBC1D4 C RAPGEF2 RAPGEF2 + + TBC1D4 DEC1 + + Cul1-Skp1-Rbx1 REST β-Catenin + – βTrCP1 –+ βTrCP1-∆F β-Catenin PDCD4

TBC1D4 (Ub)n Cul1 EDD1 250 kD -

TBC1D4 - HUWE1 150 kD Downloaded from FBXWs, CDH1, bTrCP Cul1 Fig. 6. Validation of putative SCF sub- CDC20 (Anti-FLAG) strates by coimmunoprecipitation and in vitro Skp1 ubiquitylation assay. (A) Coimmunoprecipitation of the indicated proteins with the FLAG-tagged βTrCP2 F-box proteins containing WD40 repeats http://stke.sciencemag.org/ (Anti-FLAG) (FBXWs), the APC/C activators CDH1 and CDC20 (also containing WD40 repeats), or an empty vector (EV) from HEK293T cells. Immunoprecipitation was performed with the antibody recognizing FLAG, and proteins were detected by immunoblotting with antibodies for the indicated proteins. (B) Analysis of the dependence of bTrCP-substrate interaction on the WD40 repeat of bTrCP. Proteins that coimmunoprecipated with FLAG-tagged bTrCP2WTorbTrCP2 (R447A) were analyzed by immunoblotting. (C)Invitroubiqui- tylation of TBC1D4 with SCFbTrCP1 immunoprecipitated from HEK293T cells transfected with TBC1D4, Skp1, Cul1, and Rbx1 in the presence of FLAG-tagged bTrCP1 or FLAG-tagged bTrCP1-DF. The bracket indicates a ladder of bands corresponding to polyubiquitylated TBC1D4. Data are representative of three experiments. on May 9, 2018 EDD1), both of which are components of other E3 ubiquitin ligases, had lutionary conservation to reduce the 17,779 predicted candidates to 13,109 AP-MS profiles similar to Skp1 and Cul1 (tables S3A and S4A). Therefore, sequences (6091 proteins). we tested if HUWE1 and UBR5 coimmunoprecipitated with both wild- To further limit the substrate repertoire, we used AP-MS to identify type bTrCP2 and bTrCP2 (R447A) (Fig. 6B). These results confirmed proteins that interacted with bTrCP2. To confidently discern genuine in- that both HUWE1 and UBR5 are part of the bTrCP interactome, and teractions from background contaminants, we analyzed the AP-MS data indicated that the interaction is independent of the substrate recognition using the spectra count–based CRAPome and MS1 intensity–based Per- WD40 domain. Thus, this approach not only identified substrates of an seus software tools to validate the quantitative immunoprecipitation results. E3 ubiquitin ligase but, depending on the mutants used, also revealed We found that, although there was only a modest 46% overall overlap be- potential regulators or other complexes in which the E3 ubiquitin ligase tween the CRAPome and Perseus results (Fig. 4C), the agreement was 82% may function. for the established substrates (Fig. 4D). One possible explanation for why the ability of both CRAPome and Perseus analyses performed similarly, with the highest overlap, for known substrates is the use of the bTrCP2 DISCUSSION (R447A) mutant as a negative control. The interactome for the bTrCP2 Here, we showed that searching biological sequence databases for se- (R447A) mutant theoretically resembles that of the wild-type bTrCP2, ex- quences using the MEME-FIMO software tool provides a useful initial data cept that it specifically lacks substrates. For these substrates, the AP-MS set of putative E3 ubiquitin ligase substrates by identifying, scoring, and result should represent an “on” (wild-type bTrCP2 or bTrCP2-DF) and ranking sequences that match closely to the conserved phosphodegron mo- “off ” [bTrCP2 (R447A)] situation regardless of whether spectra counts or tif. We used the DpSGXX(X)pS phosphodegron motif extracted from MS1 intensities are used to analyze the data. A newer version of the SAINT known SCFbTrCP substrates as proof of principle. This bioinformatics tool algorithm may solve the moderate overlap between the CRAPome and also captures proteins with phosphodegron that diverged in sequence from Perseus results for the interactome, because it is also based on MS1 quan- the canonical SCFbTrCP phosphodegron; these category II motifs had higher titation (25). It will be interesting to reanalyze the data once this SAINT FIMO P values. To minimize the number of possible candidates, we at- algorithm has been incorporated into CRAPome. tempted to match the short sequences to PhosphoSitePlus; however, because We further showed that only 221 of the 448 specific bTrCP interactors of the incomplete repository, the phosphorylation status of some of these contain sequences similar to the consensus motif and at the same time motifs could not be confirmed by this method. Therefore, we applied evo- were highly conserved (Fig. 4E). Therefore, not all the specific interactors

www.SCIENCESIGNALING.org 16 December 2014 Vol 7 Issue 356 rs8 7 RESEARCH RESOURCE that depended on the WD40 repeats are likely substrates; instead, they body recognizing b-catenin (BD Biosciences). Polyclonal antibodies were may represent secondary or tertiary binders. those recognizing PDCD4 (from Bethyl Laboratories), REST (Millipore), Many proteomics screens for substrates of E3 ubiquitin ligases are most- FLAG (Sigma-Aldrich), and TBC1D4 (Cell Signaling Technology). ly conducted by means of monitoring protein abundance (8–10, 13), which lack evidence pertaining to ligase-substrate interaction, because the sub- bTrCP2 immunopurification strates can be indirectly regulated, as well as regulated by ubiquitylation. HEK293T cells grown in 15-cm dishes were transfected with pcDNA3- Here, we developed an interaction proteomics approach that effectively 2xFLAG-2xHA-bTrCP2 and treated with 10 mM MG132 for 5 hours. Cells identifies most of the reported substrates of SCFbTrCP in a single screen. were harvested and subsequently lysed in lysis buffer [50 mM tris-HCl In addition, using bioinformatics, we ranked each AP-MS candidate with (pH 7.5), 150 mM NaCl, 1 mM EDTA, and 0.5% NP-40]. bTrCP2 was confidence based on the similarity of their putative degrons to the con- immunopurified with the mouse antibody against FLAG M2 coupled to sensus motif. At least four of the substrates that we identified from this an agarose resin (Sigma-Aldrich). After washing, proteins were eluted screen—eEF2K (21), RAPGEF2 (39), TFAP4 (40), and bHLHe40 (45)— by competition with FLAG peptide (Sigma-Aldrich). The eluate was then have been validated by various biochemical methods. subjected to a second immunopurification with resin linked to the anti- We expect that other baits can be designed on the basis of the substrate body against HA (12CA5 monoclonal antibody cross-linked to protein recognition substrates for other E3 ubiquitin ligases and that this combined G–Sepharose; Invitrogen). AP-MS, bioinformatic, and stringent filtering approach will lead to spe- cific and effective screens for the identification of substrates of ubiquitin Recovery of immunoprecipitated proteins for liquid ligases, especially those of the F-box family. chromatography–MS/MS Immunoprecipitated proteins were resuspended in a spin filter column (Bio-Rad) and washed three times with 200 ml of phosphate-buffered sa- Downloaded from MATERIALS AND METHODS line (pH 7.4) to remove residual detergent from the lysis buffer. Proteins bound to the protein G beads were then eluted off with 100 mlof0.5% Cell culture, transfection, and drug treatment RapiGest reagents in 50 mM ammonium bicarbonate, followed by 100 ml HEK293T cells were cultured in Dulbecco’s modified Eagle’smedium of 8 M urea in 50 mM ammonium bicarbonate (pH 8.0). Eluted proteins (Invitrogen) with 10% fetal calf serum and penicillin-streptomycin were reduced with 1 mM DTT and alkylated with 5.5 mM iodoacetamide.

(100 U/ml). HEK293T cells were transfected using the polyethylenimine For tryptic digestion, proteins were first digested with endoproteinase Lys-C http://stke.sciencemag.org/ (PEI) method and, 48 hours after transfection, treated with the proteasome (Wako Chemicals) in room temperature for 4 hours, followed by sequencing inhibitor MG132 (10 mM; Peptide Institute) for 5 hours. grade–modified trypsin (Promega) overnight, after fourfold dilution with 50 mM ammonium bicarbonate. Protease digestion was stopped by addi- DNA constructs tion of trifluoroacetic acid (TFA), and precipitates were removed after cen- Full-length, human wild-type bTrCP2 (isoform C, NP_036432.2) carrying trifugation. Peptides were desalted using reversed-phase Sep-Pak C18 HA and FLAG tags at the N terminus was cloned into the Eco RI and Not cartridges (Waters), then dried and stored at −20°C. I sites of pcDNA3 by polymerase chain reaction. The full-length construct was then used as template to generate the bTrCP2 (R447A) mutant using Liquid chromatography–MS/MS

the QuikChange Site-Directed Mutagenesis kit (Stratagene) according to For MS analysis, peptides were first separated with a C18 column (Zorbax) on May 9, 2018 the manufacturer’s directions. and introduced by nanoelectrospray into the LTQ Orbitrap Elite (Thermo The following oligonucleotides were used: forward primer, 5′- Fisher) and MS/MS in data-dependent decision tree mode (collision- GGGACATGAAGAATTGGTCGCATGCATCCGGTTTGATAACAAG-3′; induced dissociation/electron transfer dissociation) as described previously reverse primer, 5′-CTTGTTATCAAACCGGATGCATGCGACCAATTC- (51, 52). TTCATGTCCC-3′. All constructs were verified by sequencing. MS data analysis Biochemical methods MS raw files were analyzed using MaxQuant (version 1.4.0.8) with the Cells were harvested and lysed with lysis buffer [50 mM tris-HCl at pH 7.4, match between runs and LFQ options selected. MS/MS spectra, top 8 1 mM EDTA, 250 mM NaCl, 0.1% Triton X-100, 1 mM DTT (dithiothreitol)] selected in 100-dalton bin, were searched against the UniProt Human containing protease inhibitors [leupeptin (1 mg/ml), aprotinin (1 mg/ml), soy- database (version 2013-07, 20,277 entries). Trypsin/P was chosen as the bean trypsin inhibitor (10 mg/ml), 0.1 mM phenylmethylsulfonyl fluoride, protease, carbamidomethylation was set as fixed modification, tosyl lysyl chloromethyl ketone (10 mg/ml), tosylphenylalanyl chloromethyl and oxidation of methionine and acetylation of the N terminus of the protein ketone (10 mg/ml); Sigma-Aldrich] and phosphatase inhibitors (50 mM were set as variable modifications. Peptide tolerance was initially set to sodium fluoride, 100 mM sodium orthovanadate; Sigma-Aldrich) and kept 20 ppm for first search and 4.5 ppm after recalibration; MS/MS tolerance on ice for 30 min. After centrifugation at 18,500 RCF (relative centrifugal was set to 0.5 dalton. All peptide-spectrum matches (PSMs) and proteins force) in a microcentrifuge for 20 min, the supernatant was recovered and were validated with 1% false discovery rate (FDR). Only PSMs with a protein concentration was determined. Laemmli sample buffer was added minimum length of 7 amino acids were kept. to 30 mg of protein extract, which was loaded and separated by SDS– polyacrylamide gel electrophoresis. Post-acquisition data analysis: CRAPome and Perseus analyses Antibodies Before analysis using the CRAPome or Perseus software suite, the The following monoclonal antibodies were used: antibody recognizing “proteingroups.txt” table generated by MaxQuant was filtered for con- Cul1 (Invitrogen), antibody recognizing FLAG (Sigma-Aldrich), antibody taminants and highly abundant proteins (such as keratins, tubulins, and recognizing actin (Santa Cruz Biotechnology), antibody recognizing FLAG ribosomal proteins), reverse hits, number of unique peptides (>0), and num- M2 (Sigma-Aldrich), antibody recognizing HA 12CA5 (Covance), and anti- berofpeptides(>1). Subsequently, analysis and filtering using the CRAPome

www.SCIENCESIGNALING.org 16 December 2014 Vol 7 Issue 356 rs8 8 RESEARCH RESOURCE software suite were performed essentially as described (24). To discriminate SUPPLEMENTARY MATERIALS bona fide protein interactors of bTrCP2 from the background, we set a www.sciencesignaling.org/cgi/content/full/7/356/rs8/DC1 SAINT score threshold of 0.9. Subsequently, to identify the putative sub- Fig. S1. Statistical validation using the CRAPome. strates from this pool of specific interactors, we computed the ratios (FC Table S1. Reported phosphodegron motifs for bTrCP2. Table S2. List of 17,779 predicted phosphodegron motifs in human proteome. ratios) for the FC-B scores between the wild-type bTrCP2 or bTrCP2-DF Table S3. List of proteins isolated by immunoprecipitation, categorized by CRAPome against the bTrCP2 (R447A) mutant. For the convenience of FC-ratio estima- values and sorted into the reported substrates and putative substrates (data set 2). tion, missing values in FC-B scores were imputed with a minimal value of 0.1. Table S4. List of proteins isolated by immunoprecipitation, categorized by Perseus values For analysis using the Perseus software suite for identifying specific and sorted into the reported substrates and putative substrates. Table S5. List of proteins isolated by immunoprecipitation comparing wild-type bTrCP2 interactors of wild-type bTrCP2 or bTrCP2-DF against that of the bTrCP2 with R447A mutant, categorized by CRAPome values (data set 1 or old data set). (R447A) mutant, t test–based statistics were applied on LFQ (53). First, Table S6. Performance metrics of reproducibility testing. the LFQ values were transformed to logarithm (log2), and the resulting Table S7. List of putative substrates validated by both CRAPome and Perseus analyses Gaussian distribution of the data was used for imputation of missing combined with bioinformatics analyses. – values by normal distribution (width = 0.3, shift = 2.5). Statistical outliers References (58 98) were then determined using a two-tailed t test followed by multiple testing corrections with a permutation-based FDR method. REFERENCES AND NOTES 1. I. Dikic, S. Wakatsuki, K. J. Walters, Ubiquitin-binding domains—From structures to functions. Nat. Rev. Mol. Cell Biol. 10, 659–671 (2009). In vitro ubiquitylation assay 2. M. H. Glickman, A. Ciechanover, The ubiquitin-proteasome proteolytic pathway: Destruc- HEK293T cells were transfected with HA-tagged TBC1D4 along with tion for the sake of construction. Physiol. Rev. 82, 373–428 (2002). His-Skp1, Myc-Rbx1, Cul1, and either FLAG-bTrCP1 or FLAG–bTrCP1-DF 3. C. M. Pickart, M. J. Eddins, Ubiquitin: Structures, functions, mechanisms. Biochim. – using the PEI transfection method. Forty-eight hours after transfection, Biophys. Acta 1695,55 72 (2004). Downloaded from 4. C. A. M. Semple; RIKEN GER Group; GSL Members, The comparative proteomics of MG132 (10 mM) was added and cells were harvested after 5 hours. Immuno- ubiquitination in mouse. Genome Res. 13, 1389–1394 (2003). precipitation using FLAG resin was carried out as described above. Two final 5. A. W. Lau, H. Fukushima, W. Wei, The Fbw7 and beta-TRCP E3 ubiquitin ligases and – washes with QA buffer [10 mM tris-Cl (pH 7.5), 100 mM KCl, 1 mM MgCl2, their roles in tumorigenesis. Front. Biosci. 17, 2197 2212 (2012). 6. S. Lipkowitz, A. M. Weissman, RINGs of good and evil: RING finger ubiquitin ligases 0.1 mM CaCl2, 1 mM DTT] were performed in addition to the last washes with at the crossroads of tumour suppression and oncogenesis. Nat. Rev. Cancer 11, lysis buffer. After the last wash, 10 ml of the in vitro ubiquitylation mix contain- 629–643 (2011). ing 50 mM tris (pH 7.6), 5 mM MgCl2, 0.6 mM DTT, 2 mM ATP (adenosine 7. R. Fukunaga, T. Hunter, MNK1, a new MAP kinase-activated protein kinase, isolated http://stke.sciencemag.org/ 5′-triphosphate), E1 (1.5 ng/ml) (Boston Biochem), Ubc3 (10 ng/ml), ubiquitin by a novel expression screening method for identifying protein kinase substrates. (2.5 mg/ml) (Sigma-Aldrich), and 1 mM ubiquitin aldehyde was added. The re- EMBO J. 16, 1921–1933 (1997). 8. C. F. Burande, M. L. Heuzé, I. Lamsoul, B. Monsarrat, S. Uttenweiler-Joseph, P. G. Lutz, actions were incubated at 30°C for 90 min and analyzed by immunoblotting. A label-free quantitative proteomics strategy to identify E3 ubiquitin ligase substrates targeted to proteasome degradation. Mol. Cell. Proteomics 8, 1719–1727 (2009). Motif searching 9. A. Arabi, K. Ullah, R. M. M. Branca, J. Johansson, D. Bandarra, M. Haneklaus, J. Fu, The bTrCP-binding motif analysis was performed with the MEME and I. Ariës, P. Nilsson, M. L. Den Boer, K. Pokrovskaja, D. Grandér, G. Xiao, S. Rocha, FIMO tools from the MEME software suite (54) (http://meme.nbcr.net). J. Lehtiö, O. Sangfelt, Proteomic screen reveals Fbw7 as a modulator of the NF-kB pathway. Nat. Commun. 3, 976 (2012). We first identified the proper binding motif and then scanned the full human 10. B. K. Koo, M. Spit, I. Jordens, T. Y. Low, D. E. Stange, M. van de Wetering, J. H. van Es,

proteome for its presence. The motif identification was performed with S. Mohammed, A. J. R. Heck, M. M. Maurice, H. Clevers, Tumour suppressor RNF43 is a on May 9, 2018 the MEME (55) program on the set of 47 bTrCP-binding phosphorylated stem-cell E3 ligase that induces endocytosis of Wnt receptors. Nature 488,665–669 peptides publicly deposited. The motif width was set to 10 residues and (2012). 11. G. Xu, J. S. Paige, S. R. Jaffrey, Global analysis of ubiquitination by ubiquitin specified the occurrences to one per sequence for the MEME search remnant immunoaffinity profiling. Nat. Biotechnol. 28, 868–873 (2010). parameters. We then used the motif and position-specific probability matrix 12. W. Kim, E. J. Bennett, E. L. Huttlin, A. Guo, J. Li, A. Possemato, M. E. Sowa, R. Rad, generated by MEME to rescan the Swiss-Prot (human, July 2013, 20,277 J. Rush, M. J. Comb, J. W. Harper, S. P. Gygi, Systematic and quantitative assess- entries) with the FIMO (56) program. All potential binding sites with P ≤ ment of the ubiquitin-modified proteome. Mol. Cell 44, 325–340 (2011). −3 13. K. A. Lee, L. P. Hammerle, P. S. Andrews, M. P. Stokes, T. Mustelin, J. C. Silva, R. A. Black, 1×10 were reported. J. R. Doedens, Ubiquitin ligase substrate identification through quantitative proteo- mics at both the protein and peptide levels. J. Biol. Chem. 286, 41530–41538 (2011). Phosphorylation and ubiquitination sites matching 14. A. Peschiaroli, N. V. Dorrello, D. Guardavaccaro, M. Venere, T. Halazonetis, N. E. Sherman, bTrCP The phosphorylation sites and ubiquitylation sites in predicted bTrCP- M. Pagano, SCF -mediated degradation of Claspin regulates recovery from the DNA replication checkpoint response. Mol. Cell 23, 319–329 (2006). binding sites and corresponding proteins were mapped by searching the 15. N. V. Dorrello, A. Peschiaroli, D. Guardavaccaro, N. H. Colburn, N. E. Sherman, M. Pagano, PhosphoSitePlus (35) database (www.phosphosite.org; version of 2013.12.06). S6K1- and bTRCP-mediated degradation of PDCD4 promotes protein translation and cell growth. Science 314, 467–471 (2006). bTrCP-binding site conservation analysis 16. D. Guardavaccaro, D. Frescas, N. V. Dorrello, A. Peschiaroli, A. S. Multani, T. Cardozo, A. Lasorella, A. Iavarone, S. Chang, E. Hernando, M. Pagano, Control of Orthologous protein sequences and their alignments among human, mouse, – – – ’ stability by the b-TrCP REST Mad2 axis. Nature 452, 365 369 (2008). zebrafish, and fruit fly were downloaded from the EggNOG database s 17. J. W. Harper, M. K. M. Tan, Understanding -RING E3 biology through proteomics- metazoan orthologous groups (57) (http://eggnog.embl.de/version_4.0.beta/). based substrate identification. Mol. Cell. Proteomics 11, 1541–1550 (2012). The EggNOG database uses nonsupervised clustering methods to as- 18. E. Dehan, F. Bassermann, D. Guardavaccaro, G. Vasiliver-Shamis, M. Cohen, K. N. Lowes, sign 1133 species to more than 700,000 orthologous protein groups. We M. Dustin, D. C. S. Huang, J. Taunton, M. Pagano, bTrCP- and Rsk1/2-mediated deg- radation of BimEL inhibits apoptosis. Mol. Cell 33, 109–116 (2009). scanned the regions of the alignment that correspond to the bTrCP-binding 19. G. Wu, G. Xu, B. A. Schulman, P. D. Jeffrey, J. W. Harper, N. P. Pavletich, Structure of a motif (XXXDSGXXSX) in the corresponding human protein. We then as- b-TrCP1-Skp1-b-catenin complex: Destruction motif binding and lysine specificity of the sessed whether the motif observed in a human protein was also located at SCFb-TrCP1 ubiquitin ligase. Mol. Cell 11, 1445–1456 (2003). the corresponding position in any of the other three species by calculating 20. J. T. Winston, P. Strack, P. Beer-Romero, C. Y. Chu, S. J. Elledge, J. W. Harper, The SCFb-TRCP–ubiquitin ligase complex associates specifically with phosphorylated de- the percentage identity of the aligned protein sequences that corresponded struction motifs in IkBa and b-catenin and stimulates IkBa ubiquitination in vitro. to the positions of motifs. Dev. 13, 270–283 (1999).

www.SCIENCESIGNALING.org 16 December 2014 Vol 7 Issue 356 rs8 9 RESEARCH RESOURCE

21. F. Kruiswijk, L. Yuniati, R. Magliozzi, T. Y. Low, R. Lim, R. Bolder, S. Mohammed, 44. J. Cox, M. Y. Hein, C. A. Luber, I. Paron, N. Nagaraj, M. Mann, MaxLFQ allows ac- C. G. Proud, A. J. R. Heck, M. Pagano, D. Guardavaccaro, Coupled activation and curate proteome-wide label-free quantification by delayed normalization and maximal degradation of eEF2K regulates protein synthesis in response to genotoxic stress. peptide ratio extraction. Mol. Cell. Proteomics 13, 2513–2526 (2014). Sci. Signal. 5, ra40 (2012). 45. J. Kim, S. D’Annibale, R. Magliozzi, T. Y. Low, P. Jansen, I. A. Shaltiel, S. Mohammed, 22. D. Frescas, M. Pagano, Deregulated proteolysis by the F-box proteins SKP2 and b-TrCP: A. J. R. Heck, R. H. Medema, D. Guardavaccaro, USP17- and SCFbTrCP-regulated deg- Tipping the scales of cancer. Nat. Rev. Cancer 8, 438–449 (2008). radation of DEC1 controls the DNA damage response. Mol. Cell. Biol. 34, 4177–4185 23. H. Choi, B. Larsen, Z. Y. Lin, A. Breitkreutz, D. Mellacheruvu, D. Fermin, Z. S. Qin, M. Tyers, (2014). A. C. Gingras, A. I. Nesvizhskii, SAINT: Probabilistic scoring of affinity purification–mass 46. J. de Rooij, N. M. Boenink, M. van Triest, R. H. Cool, A. Wittinghofer, J. L. Bos, PDZ- spectrometry data. Nat. Methods 8,70–73 (2011). GEF1, a guanine nucleotide exchange factor specific for Rap1 and Rap2. J. Biol. Chem. 24. D. Mellacheruvu, Z. Wright, A. L. Couzens, J. P. Lambert, N. A. St-Denis, T. Li, Y. V. Miteva, 274, 38125–38130 (1999). S. Hauri, M. E. Sardiu, T. Y. Low, V. A. Halim, R. D. Bagshaw, N. C. Hubner, A. Al-Hakim, 47. Y. Liao, K. Kariya, C. D. Hu, M. Shibatohge, M. Goshima, T. Okada, Y. Watari, X. Gao, A. Bouchard, D. Faubert, D. Fermin, W. H. Dunham, M. Goudreault, Z. Y. Lin, B. G. Badillo, T. G. Jin, Y. Yamawaki-Kataoka, T. Kataoka, RA-GEF, a novel Rap1A guanine nucle- T. Pawson, D. Durocher, B. Coulombe, R. Aebersold, G. Superti-Furga, J. Colinge, otide exchange factor containing a Ras/Rap1A-associating domain, is conserved be- A. J. Heck, H. Choi, M. Gstaiger, S. Mohammed, I. M. Cristea, K. L. Bennett, M. P. Washburn, tween nematode and humans. J. Biol. Chem. 274, 37815–37820 (1999). B. Raught, R. M. Ewing, A. C. Gingras, A. I. Nesvizhskii, The CRAPome: A contaminant 48. T. Ohtsuka, Y. Hata, N. Ide, T. Yasuda, E. Inoue, T. Inoue, A. Mizoguchi, Y. Takai, nRap repository for affinity purification–mass spectrometry data. Nat. Methods 10,730–736 GEP: A novel neural GDP/GTP exchange protein for Rap1 small G protein that interacts (2013). with synaptic scaffolding molecule (S-SCAM). Biochem. Biophys. Res. Commun. 265, 25. H. Choi, T. Glatter, M. Gstaiger, A. I. Nesvizhskii, SAINT-MS1: Protein-protein inter- 38–44 (1999). action scoring using label-free intensity data in affinity purification-mass spectrometry 49. M. Larance, G. Ramm, J. Stöckli, E. M. van Dam, S. Winata, V. Wasinger, F. Simpson, experiments. J. Proteome Res. 11, 2619–2624 (2012). M. Graham, J. R. Junutula, M. Guilhaus, D. E. James, Characterization of the role of the 26. H. Choi, G. Liu, D. Mellacheruvu, M. Tyers, A. C. Gingras, A. I. Nesvizhskii, Analyzing Rab GTPase-activating protein AS160 in insulin-regulated GLUT4 trafficking. J. Biol. protein-protein interactions from affinity purification-mass spectrometry data with Chem. 280, 37803–37813 (2005). SAINT. Curr. Protoc. Bioinformatics Chapter 8, Unit 8.15 (2012). 50. C. P. Mîinea, H. Sano, S. Kane, E. Sano, M. Fukuda, J. Peränen, W. S. Lane, G. E. Lienhard, 27. J. Cox, M. Mann, MaxQuant enables high peptide identification rates, individualized AS160, the Akt substrate regulating GLUT4 translocation, has a functional Rab GTPase- Downloaded from p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat. Biotechnol. activating protein domain. Biochem. J. 391,87–93 (2005). 26, 1367–1372 (2008). 51. D. L. Swaney, G. C. McAlister, J. J. Coon, Decision tree–driven tandem mass spec- 28. J. Cox, I. Matic, M. Hilger, N. Nagaraj, M. Selbach, J. V. Olsen, M. Mann, A practical trometry for shotgun proteomics. Nat. Methods 5, 959–964 (2008). guide to the MaxQuant computational platform for SILAC-based quantitative proteomics. 52. C. K. Frese, A. F. M. Altelaar, M. L. Hennrich, D. Nolting, M. Zeller, J. Griep-Raming, Nat. Protoc. 4, 698–705 (2009). A. J. R. Heck, S. Mohammed, Improved peptide identification by targeted fragmenta- 29. J. Cox, N. Neuhauser, A. Michalski, R. A. Scheltema, J. V. Olsen, M. Mann, Andromeda: tion using CID, HCD and ETD on an LTQ-Orbitrap Velos. J. Proteome Res. 10, 2377– A peptide search engine integrated into the MaxQuant environment. J. Proteome Res. 2388 (2011).

10, 1794–1805 (2011). 53. N. C. Hubner, A. W. Bird, J. Cox, B. Splettstoesser, P. Bandilla, I. Poser, A. Hyman, http://stke.sciencemag.org/ 30. N. C. Hubner, M. Mann, Extracting function from protein–protein interactions M. Mann, Quantitative proteomics combined with BAC TransgeneOmics reveals in vivo using Quantitative BAC InteraCtomics (QUBIC). Methods 53, 453–459 (2011). protein interactions. J. Cell Biol. 189, 739–754 (2010). 31. L. Trinkle-Mulcahy, J. Andersen, Y. W. Lam, G. Moorhead, M. Mann, A. I. Lamond, 54. T. L. Bailey, M. Boden, F. A. Buske, M. Frith, C. E. Grant, L. Clementi, J. Ren, W. W. Li, Repo-Man recruits PP1g to chromatin and is essential for cell viability. J. Cell Biol. W. S. Noble, MEME SUITE: Tools for motif discovery and searching. Nucleic Acids Res. 172, 679–692 (2006). 37, W202–W208 (2009). 32. B. Blagoev, I. Kratchmarova, S. E. Ong, M. Nielsen, L. J. Foster, M. Mann, A proteomics 55. T. L. Bailey, C. Elkan, Fitting a mixture model by expectation maximization to discover strategy to elucidate functional protein-protein interactions applied to EGF signaling. motifs in biopolymers. Proc. Int. Conf. Intell. Syst. Mol. Biol. 2,28–36 (1994). Nat. Biotechnol. 21, 315–318 (2003). 56. C. E. Grant, T. L. Bailey, W. S. Noble, FIMO: Scanning for occurrences of a given motif. 33. I. Kratchmarova, B. Blagoev, M. Haack-Sorensen, M. Kassem, M. Mann, Mechanism Bioinformatics 27, 1017–1018 (2011). of divergent growth factor effects in mesenchymal stem cell differentiation. Science 57. L. J. Jensen, P. Julien, M. Kuhn, C. von Mering, J. Muller, T. Doerks, P. Bork, eggNOG: 308, 1472–1477 (2005). Automated construction and annotation of orthologous groups of genes. Nucleic Acids on May 9, 2018 34. J. A. Ranish, E. C. Yi, D. M. Leslie, S. O. Purvine, D. R. Goodlett, J. Eng, R. Aebersold, Res. 36, D250–D254 (2008). The study of macromolecular complexes by quantitative proteomics. Nat. Genet. 33, 58. Y. Li, K. G. Suresh Kumar, W. Tang, V. S. Spiegelman, S. Y. Fuchs, Negative regulation 349–355 (2003). of prolactin receptor stability and signaling mediated by SCFb-TrCP E3 ubiquitin ligase. 35. P. V. Hornbeck, J. M. Kornhauser, S. Tkachev, B. Zhang, E. Skrzypek, B. Murray, V. Latham, Mol. Cell. Biol. 24, 4038–4048 (2004). M. Sullivan, PhosphoSitePlus: A comprehensive resource for investigating the structure 59. B. P. Zhou, J. Deng, W. Xia, J. Xu, Y. M. Li, M. Gunduz, M. C. Hung, Dual regulation of and function of experimentally determined post-translational modifications in man and Snail by GSK-3b-mediated phosphorylation in control of epithelial–mesenchymal tran- mouse. Nucleic Acids Res. 40, D261–D270 (2012). sition. Nat. Cell Biol. 6, 931–940 (2004). 36. R. Gutman, C. Berezin, R. Wollman, Y. Rosenberg, N. Ben-Tal, QuasiMotiFinder: Protein 60. I. Lassot, E. Ségéral, C. Berlioz-Torrent, H. Durand, L. Groussin, T. Hai, R. Benarous, annotation by searching for evolutionarily conserved motif-like patterns. Nucleic Acids F. Margottin-Goguet, ATF4 degradation relies on a phosphorylation-dependent inter- Res. 33, W255–W261 (2005). action with the SCFbTrCP ubiquitin ligase. Mol. Cell. Biol. 21, 2192–2202 (2001). 37. N. E. Davey, D. C. Shields, R. J. Edwards, Masking residues using context-specific evo- 61. J. Pons, N. Evrard-Todeschi, G. Bertho, J. Gharbi-Benarous, V. Tanchou, R. Benarous, lutionary conservation significantly improves short linear motif discovery. Bioinformatics J. P. Girault, Transfer-NMR and docking studies identify the binding of the peptide 25, 443–450 (2009). derived from activating factor 4 to protein ubiquitin ligase b-TrCP. Competi- 38. M. Peng, A. Scholten, A. J. R. Heck, B. van Breukelen, Identification of enriched PTM cross- tion STD-NMR with b-catenin. Biochemistry 47,14–29 (2008). talk motifs from large-scale experimental data sets. J. Proteome Res. 13,249–259 (2014). 62. H. Chen, H. Ma, H. Inuzuka, J. Diao, F. Lan, Y. G. Shi, W. Wei, Y. Shi, DNA damage 39. R. Magliozzi, T. Y. Low, B. G. M. W. Weijts, T. Cheng, E. Spanjaard, S. Mohammed, regulates UHRF1 stability via the SCFb-TrCP E3 ligase. Mol. Cell. Biol. 33, 1139–1148 A. van Veen, H. Ovaa, J. de Rooij, F. J. T. Zwartkruis, J. L. Bos, A. de Bruin, A. J. R. Heck, (2013). D. Guardavaccaro, Control of epithelial cell migration and invasion by the IKKb-and 63. D. V. Hansen, A. V. Loktev, K. H. Ban, P. K. Jackson, Plk1 regulates activation of the CK1a-mediated degradation of RAPGEF2. Dev. Cell 27, 574–585 (2013). anaphase promoting complex by phosphorylating and triggering SCFbTrCP-dependent 40. S. D’Annibale,J.Kim,R.Magliozzi,T.Y.Low,S.Mohammed,A.J.R.Heck,D.Guardavaccaro, destruction of the APC Inhibitor Emi1. Mol. Biol. Cell 15, 5623–5634 (2004). Proteasome-dependent degradation of transcription factor activating enhancer-binding 64. J. Jin, T. Shirogane, L. Xu, G. Nalepa, J. Qin, S. J. Elledge, J. W. Harper, SCFb-TRCP protein 4 (TFAP4) controls mitotic division. J. Biol. Chem. 289, 7730–7737 (2014). links Chk1 signaling to degradation of the Cdc25A protein phosphatase. Genes Dev. 41. A. Yaron, A. Hatzubai, M. Davis, I. Lavon, S. Amit, A. M. Manning, J. S. Andersen, M. Mann, 17, 3062–3074 (2003). F. Mercurio, Y. Ben-Neriah, Identification of the receptor component of the IkBa-ubiquitin 65. R. E. Amir, H. Haecker, M. Karin, A. Ciechanover, Mechanism of processing of the ligase. Nature 396, 590–594 (1998). NF-kB2 p100 precursor: Identification of the specific polyubiquitin chain-anchoring 42. F. Margottin, S. P. Bour, H. Durand, L. Selig, S. Benichou, V. Richard, D. Thomas, K. Strebel, lysine residue and analysis of the role of NEDD8-modification on the SCFb-TrCP ubiquitin R. Benarous, A novel human WD protein, h-bTrCP, that interacts with HIV-1 Vpu connects ligase. Oncogene 23, 2540–2547 (2004). CD4 to the ER degradation pathway through an F-box motif. Mol. Cell 1, 565–574 (1998). 66. Y. Kee, J. M. Kim, A. D. D’Andrea, Regulated degradation of FANCM in the Fanconi 43. J. Cox, M. Mann, 1D and 2D annotation enrichment: A statistical method integrating anemia pathway during mitosis. Genes Dev. 23, 555–560 (2009). quantitative proteomics with complementary high-throughput data. BMC Bioinformatics 67. B. Zhao, L. Li, K. Tumaneng, C. Y. Wang, K. L. Guan, A coordinated phosphorylation by 13 (Suppl. 1), S12 (2012). Lats and CK1 regulates YAP stability through SCFb-TRCP. Genes Dev. 24,72–85 (2010).

www.SCIENCESIGNALING.org 16 December 2014 Vol 7 Issue 356 rs8 10 RESEARCH RESOURCE

68. T. Shirogane, J. Jin, X. L. Ang, J. W. Harper, SCFb-TRCP controls clock-dependent tran- 88. G. Guderian, J. Westendorf, A. Uldschmid, E. A. Nigg, Plk4 trans-autophosphorylation scription via casein kinase 1-dependent degradation of the mammalian period-1 (Per1) regulates centriole number by controlling bTrCP-mediated degradation. J. Cell Sci. 123, protein. J. Biol. Chem. 280, 26863–26872 (2005). 2163–2169 (2010). 69. S. Chowdhry, Y. Zhang, M. McMahon, C. Sutherland, A. Cuadrado, J. D. Hayes, Nrf2 is 89. H. Fukushima, K. Ogura, L. Wan, Y. Lu, V. Li, D. Gao, P. Liu, A. W. Lau, T. Wu, M. W. Kirschner, controlled by two distinct b-TrCP recognition motifs in its Neh6 domain, one of which can H. Inuzuka, W. Wei, SCF-mediated Cdh1 degradation defines a negative feedback sys- be modulated by GSK-3 activity. Oncogene 32, 3765–3781 (2013). tem that coordinates cell-cycle progression. Cell Rep. 4, 803–816 (2013). 70. M. Hayakawa, H. Kitagawa, K. Miyazawa, M. Kitagawa, K. Kikugawa, The FWD1/b-TrCP- 90. N. Popov, C. Schülein, L. A. Jaenicke, M. Eilers, Ubiquitylation of the amino terminus of mediated degradation pathway establishes a “turning off switch” of a Cdc42 guanine nu- Myc by SCFb-TrCP antagonizes SCFFbw7-mediated turnover. Nat. Cell Biol. 12, 973–981 cleotide exchange factor, FGD1. Genes Cells 10, 241–251 (2005). (2010). 71. K. Ohsaki, K. Oishi, Y. Kozono, K. Nakayama, K. I. Nakayama, N. Ishida, The role of 91. M. L. Li, J. Defren, G. Brewer, Hsp27 and F-box protein b-TrCP promote degradation b-TrCP1 and b-TrCP2 in circadian rhythm generation by mediating degradation of of mRNA decay factor AUF1. Mol. Cell. Biol. 33, 2315–2326 (2013). clock protein PER2. J. Biochem. 144, 609–618 (2008). 92. Y. Kanemori, K. Uto, N. Sagata, b-TrCP recognizes a previously undescribed nonpho- 72. S. Reischl, K. Vanselow, P. O. Westermark, N. Thierfelder, B. Maier, H. Herzel, A. Kramer, sphorylated destruction motif in Cdc25A and Cdc25B phosphatases. Proc. Natl. b-TrCP1-mediated degradation of PERIOD2 is essential for circadian dynamics. J. Biol. Acad. Sci. U.S.A. 102, 6279–6284 (2005). Rhythms 22, 375–386 (2007). 93. H. Inuzuka, A. Tseng, D. Gao, B. Zhai, Q. Zhang, S. Shaik, L. Wan, X. L. Ang, C. Mock, 73. A. C. da Silva Almeida, G. J. Strous, A. G. S. H. van Rossum, bTrCP controls GH H. Yin, J. M. Stommel, S. Gygi, G. Lahav, J. Asara, Z. X. J. Xiao, W. G. Kaelin, J. W. Harper, receptor degradation via two different motifs. Mol. Endocrinol. 26, 165–177 (2012). W. Wei, Phosphorylation by casein kinase I promotes the turnover of the Mdm2 oncoprotein 74. Y. Zhao, X. Xiong, Y. Sun, DEPTOR, an mTOR inhibitor, is a physiological substrate via the SCFb-TRCP ubiquitin ligase. Cancer Cell 18,147–159 (2010). of SCFbTrCP E3 ubiquitin ligase and regulates survival and autophagy. Mol. Cell 44, 94. J. Zhong, K. Ogura, Z. Wang, H. Inuzuka, Degradation of the transcription factor Twist, 304–316 (2011). an oncoprotein that promotes cancer metastasis. Discov. Med. 15,7–15 (2013). 75. A. C. Burrows, J. Prokop, M. K. Summers, Skp1-Cul1-F-box ubiquitin ligase (SCFbTrCP)- 95. Q. Ding, X. He, J. M. Hsu, W. Xia, C. T. Chen, L. Y. Li, D. F. Lee, J. C. Liu, Q. Zhong,

mediated destruction of the ubiquitin-specific protease USP37 during G2-phase promotes X. Wang, M. C. Hung, Degradation of Mcl-1 by b-TrCP mediates glycogen synthase mitotic entry. J. Biol. Chem. 287, 39021–39029 (2012). kinase 3-induced tumor suppression and chemosensitization. Mol. Cell. Biol. 27, 76. K. G. S. Kumar, W. Tang, A. K. Ravindranath, W. A. Clark, E. Croze, S. Y. Fuchs, 4006–4017 (2007). Downloaded from SCFHOS ubiquitin ligase mediates the ligand-induced down-regulation of the interferon-a 96. P. C. Chu, H. C. Chuang, S. K. Kulp, C. S. Chen, The mRNA-stabilizing factor HuR receptor. EMBO J. 22, 5480–5490 (2003). protein is targeted by b-TrCP protein for degradation in response to glycolysis inhibition. 77. N. Watanabe, H. Arai, Y. Nishihara, M. Taniguchi, N. Watanabe, T. Hunter, H. Osada, J. Biol. Chem. 287, 43639–43650 (2012). M-phase kinases induce phospho-dependent ubiquitination of somatic Wee1 by 97. E. Estrabaud, I. Lassot, G. Blot, E. Le Rouzic, V. Tanchou, E. Quemeneur, L. Daviet, SCFb-TrCP. Proc. Natl. Acad. Sci. U.S.A. 101, 4419–4424 (2004). F. Margottin-Goguet, R. Benarous, RASSF1C, an isoform of the tumor suppressor 78. S. Wei, H. C. Chuang, W. C. Tsai, H. C. Yang, S. R. Ho, A. J. Paterson, S. K. Kulp, C. S. Chen, RASSF1A, promotes the accumulation of b-catenin by interacting with bTrCP. Cancer Thiazolidinediones mimic glucose starvation in facilitating Sp1 degradation through the up- Res. 67, 1054–1061 (2007).

regulation of b-transducin repeat-containing protein. Mol. Pharmacol. 76,47–57 (2009). 98. G. Coadou, J. Gharbi-Benarous, S. Megy, G. Bertho, N. Evrard-Todeschi, E. Segeral, http://stke.sciencemag.org/ 79.S.Hatakeyama,M.Kitagawa,K.Nakayama,M.Shirane,M.Matsumoto,K.Hattori, R. Benarous, J. P. Girault, NMR studies of the phosphorylation motif of the HIV-1 protein H. Higashi, H. Nakano, K. Okumura, K. Onoé, R. A. Good, Ubiquitin-dependent degra- Vpu bound to the F-box protein b-TrCP. Biochemistry 42, 14741–14751 (2003). dation of IkBa is mediated by a ubiquitin ligase Skp1/Cul 1/F-box protein FWD1. Proc. Natl. Acad. Sci. U.S.A. 96, 3859–3863 (1999). Funding: This work has been supported by the Netherlands Proteomics Centre, the Nether- 80. A. Seki, J. A. Coppinger, H. Du, C. Y. Jang, J. R. Yates III, G. Fang, Plk1- and b-TrCP– lands Organization for Scientific Research (NWO) supporting the Roadmap-embedded dependent degradation of Bora controls mitotic progression. J. Cell Biol. 181,65–78 (2008). large-scale proteomics facility Proteins@Work (project 184.032.201), and the PRIME-XS 81. W. Huang, X. Lv, C. Liu, Z. Zha, H. Zhang, Y. Jiang, Y. Xiong, Q. Y. Lei, K. L. Guan, The project grant agreement number 262067 supported by the European Community’s Sev- N-terminal phosphodegron targets TAZ/WWTR1 protein for SCFb-TrCP-dependent enth Framework Programme (FP7/2007-2013) to A.J.R.H. Author contributions: T.Y.L. degradation in response to phosphatidylinositol 3-kinase inhibition. J. Biol. Chem. 287, designed, performed, and analyzed the proteomics experiments; M.P. devised and opti- 26245–26253 (2012). mized the bioinformatics workflow and analyzed the proteomics data; R.M. created and 82. C. Y. Liu, Z. Y. Zha, X. Zhou, H. Zhang, W. Huang, D. Zhao, T. Li, S. W. Chan, C. J. Lim, optimized the different bait constructs and transfected cell lines, and performed coim- on May 9, 2018 W. Hong, S. Zhao, Y. Xiong, Q. Y. Lei, K. L. Guan, The Hippo tumor pathway promotes munoprecipitation and all biochemical validations. T.Y.L., D.G., and A.J.R.H. conceived TAZ degradation by phosphorylating a phosphodegron and recruiting the SCFb-TrCP E3 the idea for this study. S.M., D.G., and A.J.R.H. supervised the study and co-wrote the ligase. J. Biol. Chem. 285, 37159–37169 (2010). manuscript with T.Y.L. Competing interests: The authors declare that they have no 83. L. Meyer, B. Deau, H. Forejtníková, D. Duménil, F. Margottin-Goguet, C. Lacombe, P. Mayeux, competing financial interests. Data and materials availability: The MS proteomics data F. Verdier, b-Trcp mediates ubiquitination and degradation of the erythropoietin receptor (data set 1) have been deposited to the ProteomeXchange Consortium via the PRIDE and controls cell proliferation. Blood 109, 5215–5222 (2007). partner repository with the data set identifier PXD001088. The reanalyzed GeLC-MS 84. V. Lang, J. Janzen, G. Z. Fischer, Y. Soneji, S. Beinke, A. Salmeron, H. Allen, R. T. Hay, data set (data set 2) for reproducibility evaluation is available with the data set identifier Y. Ben-Neriah, S. C. Ley, bTrCP-mediated proteolysis of NF-kB1 p105 requires phos- PXD001224. phorylation of p105 927 and 932. Mol. Cell. Biol. 23, 402–413 (2003). 85. F. Mantovani, L. Banks, Regulation of the discs large tumor suppressor by a phosphorylation- dependent interaction with the b-TrCP ubiquitin ligase receptor. J. Biol. Chem. 278, Submitted 8 September 2014 42477–42486 (2003). Accepted 21 November 2014 86. D. Setoyama, M. Yamashita, N. Sagata, Mechanism of degradation of CPEB during Final Publication 16 December 2014 oocyte maturation. Proc. Natl. Acad. Sci. U.S.A. 104, 18001–18006 (2007). 10.1126/scisignal.2005882 87. X. L. Ang, D. P. Seeburg, M. Sheng, J. W. Harper, Regulation of postsynaptic RapGAP Citation: T. Y. Low, M. Peng, R. Magliozzi, S. Mohammed, D. Guardavaccaro, A. J. R. Heck, A SPAR by Polo-like kinase 2 and the SCFb-TRCP ubiquitin ligase in hippocampal neurons. systems-wide screen identifies substrates of the SCFbTrCP ubiquitin ligase. Sci. Signal. 7,rs8 J. Biol. Chem. 283, 29424–29432 (2008). (2014).

www.SCIENCESIGNALING.org 16 December 2014 Vol 7 Issue 356 rs8 11 A systems-wide screen identifies substrates of the SCFβTrCP ubiquitin ligase Teck Yew Low, Mao Peng, Roberto Magliozzi, Shabaz Mohammed, Daniele Guardavaccaro and Albert J. R. Heck

Sci. Signal. 7 (356), rs8. DOI: 10.1126/scisignal.2005882

Looking for Substrates in the Proteomic Haystack System-wide analysis of protein-protein interactions is challenging because there can be many false positives in the data. Here, Low et al. describe a method not only for identifying the interacting partners of an enzyme that targets proteins for degradation but also for sorting the interactome into high-confidence pools of potential substrates and regulators. The method relies on the ability to define a consensus motif in the substrates, as well as knowledge of the region of the enzyme subunit responsible for recognizing this motif. Although the authors applied this method only to the ubiquitin ligase SCFβTrCP, this approach should be applicable to other ubiquitin ligases, especially those that use F-box Downloaded from proteins other than βTrCP to recognize its substrates.

ARTICLE TOOLS http://stke.sciencemag.org/content/7/356/rs8 http://stke.sciencemag.org/

SUPPLEMENTARY http://stke.sciencemag.org/content/suppl/2014/12/12/7.356.rs8.DC1 MATERIALS

RELATED http://stke.sciencemag.org/content/sigtrans/4/197/ra73.full CONTENT http://stke.sciencemag.org/content/sigtrans/5/219/ra30.full http://stke.sciencemag.org/content/sigtrans/9/421/ec73.abstract on May 9, 2018

REFERENCES This article cites 98 articles, 44 of which you can access for free http://stke.sciencemag.org/content/7/356/rs8#BIBL

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Use of this article is subject to the Terms of Service

Science Signaling (ISSN 1937-9145) is published by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title Science Signaling is a registered trademark of AAAS.